基于决策支持度的决策树生成算法  

Decision trees generation algorithm based on decision support degree

作　　者：关晓蔷[1] 梁吉业[1] 钱宇华[1] 刘煜伟 

机构地区：[1]山西大学计算机与信息技术学院,太原030006 [2]山西省军区司令部,太原030013

出　　处：《计算机工程与应用》2008年第27期148-150,共3页Computer Engineering and Applications

基　　金：国家自然科学基金(No.70471003);高等院校博士学科点专项科研基金(No.20050108004);教育部科学技术研究重点项目(No.206017);山西省青年科技研究基金(No.2006021019)~~

摘　　要：从条件属性对决策支持程度不同的角度出发,引入了决策支持度的概念,提出了一种以其为启发式信息的决策树生成算法。实验分析表明,相对于传统的决策树生成算法,此算法改善了决策树的结构,有效提高了决策分类的精度。Based on the viewpoint that conditions attributes have different decision support ability,the concept of decision support degree is introduced,and a novel algorithm for building decision tree is proposed, in which the decision support degree is regarded as heuristic information.The experimental analyzes on several public date sets show that unlike the decision tree built by traditional algorithms,the decision tree built by this algorithm has much better tree structure and classification precision.

关 键 词：决策树 信息系统 决策支持度 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]